Turn Your AI Assistant Into a GIS Admin

Stop wrestling with GeoServer's REST API documentation. This MCP server lets Claude, Cursor, or any AI assistant manage your geospatial infrastructure through natural language commands.

The Problem with GeoServer Management

If you've worked with GeoServer, you know the drill. Need to create a workspace? Navigate through the admin interface or craft the exact REST call. Want to query spatial features? Hope you remember the CQL syntax. Publishing a new layer? Better have the JSON payload structure memorized.

These repetitive tasks eat up development time, especially when you're prototyping spatial applications or managing multiple GeoServer instances.

What geoserver-mcp Changes

Instead of this:

curl -u admin:geoserver -XPOST -H "Content-type: text/xml" \
  http://localhost:8080/geoserver/rest/workspaces \
  -d "<workspace><name>my_workspace</name></workspace>"

You simply tell your AI assistant: "Create a workspace called my_workspace"

The MCP server handles the REST API complexity while your assistant provides the natural language interface you actually want to use.

Real Productivity Gains

Spatial Data Exploration: Ask "Show me all states with population over 10 million" instead of writing CQL queries by hand. Your assistant executes the spatial query and returns results.

Layer Management: "Create a new layer for storing building footprints with these attributes..." Your assistant handles the GeoServer configuration, field definitions, and SRS setup.

Map Generation: "Generate a styled map of the western US showing population density" becomes a single command instead of configuring WMS parameters, bounding boxes, and styling.

Development Workflow: During development, you can rapidly prototype spatial features by describing what you need rather than context-switching to GeoServer documentation.

Installation That Actually Works

Three installation paths depending on your setup:

Docker (recommended for testing):

docker pull mahdin75/geoserver-mcp
docker run -d -e GEOSERVER_URL=http://localhost:8080/geoserver \
  -e GEOSERVER_USER=admin -e GEOSERVER_PASSWORD=geoserver \
  mahdin75/geoserver-mcp

pip (for regular use):

uv venv --python=3.10
uv pip install geoserver-mcp
export GEOSERVER_URL="http://localhost:8080/geoserver"
geoserver-mcp

Development (for contributors): Clone, install with uv pip install -e ., and you're ready to modify the codebase.

What Your Assistant Can Do

The server exposes 15+ tools covering the GeoServer operations you actually use:

• Data Management: Create workspaces, manage layers, publish rasters
• Spatial Queries: Execute CQL filters, query features by attributes or geometry
• Visualization: Generate maps with custom styling, apply SLD styles
• OGC Services: Access WMS/WFS capabilities, work with standards-compliant endpoints

Each tool is designed around real workflows - not just API coverage for the sake of completeness.

Integration with Your Workflow

Configure once in your MCP client (Claude Desktop or Cursor):

{
  "mcpServers": {
    "geoserver-mcp": {
      "command": "/path/to/.venv/bin/geoserver-mcp",
      "args": ["--url", "http://localhost:8080/geoserver", "--user", "admin", "--password", "geoserver"]
    }
  }
}

Then your spatial data infrastructure becomes as accessible as asking questions. No more keeping GeoServer docs open in another tab.

Perfect for Spatial Developers

This MCP server shines when you're:

• Prototyping spatial applications and need to quickly set up test data
• Managing multiple GeoServer instances across development/staging/production
• Exploring spatial datasets without writing custom scripts for every query
• Teaching or demonstrating GIS concepts without getting lost in interface complexity
• Automating spatial workflows through conversational interfaces

At 22 stars and active development, geoserver-mcp is building momentum in the spatial developer community. The alpha status means it's ready for testing and experimentation, with the core functionality solid enough for real development work.

If you're already using AI assistants for development and work with geospatial data, this MCP server belongs in your toolkit. It's the kind of productivity multiplier that makes you wonder how you managed spatial infrastructure before having an AI assistant that actually understands GeoServer.